Three decades of research in the United States indicates that Americans who are dying often spend their final days in pain and receiving treatments they would not choose. One in five deaths among adults occurs in or shortly after discharge from an intensive care unit, where there are well-documented problems with clinician- family communication and end-of-life care. These deficiencies are associated with high rates of adverse bereavement outcomes among family members and poor quality end-of-life care. Unfortunately, there are no practical, generalizable interventions proven effective in mitigating these public health problems. In a pilot project, we developed the PARTNER intervention (PAiring Re-engineered ICU Teams with Nurse- driven Emotional Support and Relationship-building), an interdisciplinary intervention that 1) gives new responsibilities and advanced communication skills training to the local nurse leaders and social work members of the ICU team; 2) changes care defaults to ensure frequent clinician-family meetings; 3) adds protocolized, nurse-administered coaching and emotional support of surrogates before and during clinician- family meetings, and 4) increases use of palliative care services for patients with a poor prognosis. Our 2-year pilot study documented that the intervention is feasible, sustainable, and is associated with shorter ICU length of stay and lower hospital costs. The objective of this proposal is to conduct a multi-center, randomized stepped wedge trial testing the PARTNER intervention in 5 ICUs among 1000 patients with advanced critical illness and their surrogates. In Aim 1, we will assess the intervention's impact on surrogates' symptoms of anxiety, depression, PTSD, and decisional regret over 6 months of follow-up. In Aim 2, we will use validated instruments to assess the intervention's impact on communication about end-of-life decisions and patients' end-of-life experiences. In Aim 3, we will assess the intervention's impact on health care utilization during the index hospitalization and over 6 months of follow-up. The PARTNER intervention will have high impact if successful because it will be a pragmatic solution to important public health problems that potentially affect more than 600,000 adults who die annually in ICUs. The intervention is innovative in its theoretical grounding in decision psychology and behavioral economics, the breadth and intensity of support provided to surrogates, the systems-level design of the intervention, and the strategy to use the existing clinical team to deploy the intervention instead of hiring external interventionists. The work is feasible in our hands because our team of established investigators developed and successfully pilot-tested the intervention, and because we have a proven record of success conducting multi-center trials in ICUs.